Bizarre - Stata changing variable values without explanation

[Richard Hall]

I've been using Stata for 30 years and never have I come across a glitch like this. I check the simple statistics for a variable. Correlate that variable with two other variables, drop a couple of varialbes, then check the simple statistics on the first variable again. The values on it have changed by an order of magnitude! See attached pdf.

[Jared Greathouse]

I wanna see the exact code you used. Post it, please. Along with an example dataset that illustrates the problem using dataex. (I nor some others will download a dataset from someone they don't know, so please, use dataex).

[Tiago Pereira]

Richard, do you have access to an original version of Stata 17.0/SE?

[Nick Cox]

The code used is shown in the attachment but it doesn't explain the problem for me. I agree: This looks bizarre and I can offer only the useless explanation that either the dataset or the software is corrupted in some way. The dataset looks too big to post here.